def predictions(dF): """ features = dF[['rain']] dummy_units = pd.get_dummies(dF['UNIT'], prefix='unit') dummy_hour = pd.get_dummies(dF['hour'], prefix='hour') dummy_daywk = pd.get_dummies(dF['day_week'], prefix='day_week') features = features.join(dummy_units).join(dummy_hour).join(dummy_daywk) #Values values = dF['ENTRIESn_hourly'] """ features, values = gad(dF) #Get arrays features_arr = features.values values_arr = values.values means, std_devs, normalized_features_arr = normalize_feature(features_arr) #Perform linear regression withgradient descent norm_intercept, norm_params = linear_regression(normalized_features_arr, values_arr) intercept , params = recover_params(means, std_devs, norm_intercept, norm_params) print len(params) print 'Coefficient of non dummy variable by Gradient Descent is ', params[0] predictions = intercept + np.dot(features_arr, params) return predictions
def predictions(dF): """ #Features features = dF[['rain']] dummy_units = pd.get_dummies(dF['UNIT'], prefix='unit') dummy_hour = pd.get_dummies(dF['hour'], prefix='hour') dummy_daywk = pd.get_dummies(dF['day_week'], prefix='day_week') features = features.join(dummy_units).join(dummy_hour).join(dummy_daywk) #Values values = dF['ENTRIESn_hourly'] """ features, values = gad(dF) #Perform linear regression intercept, params = linear_regression(features, values) print len(params) predictions = intercept + np.dot(features, params) return predictions